Testing cookie behavior across hundreds of web surfaces with our in-house auditor
Our cookie auditor acts like a privacy-conscious user by visiting web pages and checking that they only load cookies consistent with that user’s preferences.
Improving infrastructure efficiency for growing demand in the age of AI
As demand for AI continues to grow, so does the infrastructure needed to support it.
SMR: What we learned in our first year
How our universal content processing platform Riviera evolved for AI and beyond
Riviera is the Dropbox content processing platform that’s been iteratively improving content transformation in our products for roughly a decade.
How we used DSPy to turn AI evaluations into better responses in Dash chat
We used DSPy to improve LLM judges and optimize our chat experience, creating an evaluation-driven feedback loop that produced better outputs.
How Dropbox uses MCP and Dash to close the design-to-code security gap
Using an agentic AI system to surface threat models during code review and spot gaps between security requirements and implementation.
Beyond code generation: rethinking engineering productivity in the age of AI agents
How Dropbox is moving from AI tools that assist engineers to agentic systems that can execute scoped tasks, and how we’re building platforms to support those workflows.
Introducing Nova, our internal platform for coding agents
Nova lets engineers run multiple coding sessions in parallel and lets internal systems use AI agents as part of automated workflows.
How we optimized Dash's relevance judge with DSPy
We used DSPy to turn prompt engineering for our relevance judge into a measurable, automated optimization loop, improving task performance, cost, and how reliably it works in production.
Reducing our monorepo size to improve developer velocity
Monorepos will continue to grow as products evolve, but growth doesn’t have to mean friction.
Improving storage efficiency in Magic Pocket, our immutable blob store
By turning compaction into a layered, adaptive pipeline and strengthening our monitoring and controls, we made Magic Pocket more resilient to workload changes.
Using LLMs to amplify human labeling and improve Dash search relevance
How we train Dash's search ranking models with a mix of human and LLM-assisted labeling.
How low-bit inference enables efficient AI
Making products like Dropbox Dash accessible to individuals and businesses means tackling new challenges around efficiency and resource use.
Engineering VP Josh Clemm on how we use knowledge graphs, MCP, and DSPy in Dash
Engineering VP Josh Clemm deep-dives into how we think about knowledge graphs, indexes, MCP, and prompt optimization using tools like DSPy.
Inside the feature store powering real-time AI in Dropbox Dash
The feature store is a critical part of how we rank and retrieve the right context across your work.
How Dash uses context engineering for smarter AI
Building effective, agentic AI isn’t just about adding more; it’s about helping the model focus on what matters most.
With Mobius Labs' Aana models, we're bringing deeper multimodal understanding to Dropbox Dash
Dropbox welcomes Mobius Labs to advance Dash’s multimodal AI, integrating Aana’s efficient architecture to enhance photo and video understanding at Dropbox scale.
Half-Quadratic Quantization of large machine learning models
Learn how Half-Quadratic Quantization (HQQ) makes it easy to compress large AI models without sacrificing accuracy—no calibration data required.
How we reduced the size of our JavaScript bundles by 33%
Building Dash: How RAG and AI agents help us meet the needs of businesses
To meet the challenges of modern work in data-intensive environments, we ultimately turned to retrieval-augmented generation and AI agents.
A practical blueprint for evaluating conversational AI at scale
Building Dropbox Dash taught us that in the foundation-model era, AI evaluations matter just as much as model training.
How we use Lakera Guard to secure our LLMs
With its robust capabilities, Lakera Guard helps us secure and protect user data, and uphold the reliability and trustworthiness of our intelligent features.
Making file encryption fast and secure for teams with advanced key management
We developed features to help teams limit their security risks and respond more effectively to potential threats or breaches.
Evolving our infrastructure through the messaging system model in Dropbox
The MSM helped us build a unified event-driven system capable of orchestrating a wide range of asynchronous tasks and meeting future needs, especially as we focus on AI.
How we brought multimedia search to Dropbox Dash
Our multimedia retrieval features allow users to find images, video, and audio just as easily as they find documents.
Seventh-generation server hardware at Dropbox: our most efficient and capable architecture yet
This generation represents our most efficient, capable, and scalable architecture yet—and it’ll help us as we continue to build AI products like Dropbox Dash.
Our journey to type checking 4 million lines of Python
Why we built a custom Rust library for Capture
How Edison is helping us build a faster, more powerful Dropbox on the web
Fast and Accurate Document Detection for Scanning
Cannes: How ML saves us $1.7M a year on document previews
How image search works at Dropbox
Optimizing payments with machine learning
Accelerating our A/B experiments with machine learning
Don't you (forget NLP): Prompt injection with control characters in ChatGPT
Is this a date? Using ML to identify date formats in file names
Putting everything in its right place with ML-powered file organization
Smart move uses machine learning to analyze a user’s existing subfolder structure and suggest folders where they might want to move their files. Here's how the feature was built.
Bye Bye Bye...: Evolution of repeated token attacks on ChatGPT models
Building on prior prompt injection research, we recently discovered a new training data extraction vulnerability involving OpenAI’s chat completion models.
Bringing AI-powered answers and summaries to file previews on the web
Selecting a model for semantic search at Dropbox scale
Nautilus is our search engine for finding documents and other files in Dropbox. Here's how it created the foundation for us to build better search functionality to understand more nuanced queries.
zxcvbn: realistic password strength estimation
Implementing end-to-end encryption for Dropbox teams
Learn more about our implementation of end-to-end encryption for teams, the threat model of our design and encryption algorithms, and our commitment to minimizing the risk of data loss with a team-centric key management approach.
Inside the Magic Pocket
Optimizing web servers for high throughput and low latency
Rewriting the heart of our sync engine
Testing sync at Dropbox
How we migrated Dropbox from Nginx to Envoy
Broccoli: Syncing faster by syncing less
How we designed Dropbox’s ATF - an async task framework
Alki, or how we learned to stop worrying and love cold metadata
Lessons learned in incident management
Boosting Dropbox upload speed and improving Windows’ TCP stack
That time we unplugged a data center to test our disaster readiness
A look inside our sixth generation of server hardware
Everything in its write place: Cloud storage abstraction with Object Store
Future-proofing our metadata stack with Panda, a scalable key-value store
Increasing Magic Pocket write throughput by removing our SSD cache disks
Balancing quality and coverage with our data validation framework
After four years of SMR storage, here's what we love—and what comes next
